Você está na página 1de 10

A DIFFERENT LOOK AT THE WELFARE TRAP: INSTITUTIONAL

CAUSES AND REMEDIES


Necati Celik
Abstract: This paper investigates the institutional nexus of Welfare Support Programs
(WSPs) in the US and households access to social needs, particularly affordable
healthcare and education. A close look at the Survey of Consumer Finances (SCF) reveals
that households that were turned down when they applied for credit in the past are
more likely to receive welfare support in the survey year. Moreover, welfare recipient
households report lower status of health and lower years of education regardless of
their income and demographic attributes. These results support the hypothesis that
WSPs limit households access to healthcare and education due to the institutional
set-up of the two in the US. Constrained-households find themselves in the welfare
trap because their limited access to better health care and education are reflected in
their household income, pushing them further down the ladder and increasing their
dependency on WSPs for subsistence. Neither the repeal nor the expansion of WSPs
will solve this dilemma unless the US government enacts institutional policies to
address the way higher education and health-care is provided and financed in the
country. Affordability of health care and higher education should be the main target
of these policies to enable households to pull themselves out of the welfare trap by
finding and maintaining jobs without regressing back in to welfare support.
Keywords: welfare trap, welfare programs, credit rationing, health, U.S. households
JEL Classification Codes: I380, I10

Welfare Trap is only one of the many contradictions of welfare policies. The term
denotes a welfare recipients reluctance and/or inability to break his/her dependence
on public assistance. In a situation like this, welfare support perpetuates the dependency
of the recipient on public assistance rather than raising them to a level that they dont
need assistance anymore. There are two different views on why recipients can get stuck
in welfare trap (Fraser & Gordon 1994). The first one takes a behavioral approach and
explains that welfare trap is caused by recipients lack of necessary traits to climb the
socio-economic ladder. Lack of these traits might be associated with individual
psychology or social environment. According to this view, being in welfare trap is a
result of the mistakes recipients make over and over again in their decisions. The other
view is a direct reflection of economic orthodoxy and suggests that welfare trap is a
result of an economic optimization individuals make when they face the choices of
receiving benefits or joining the workforce. Individuals deliberately choose to receive
benefits instead of finding employment because the cost of working is too high
compared to the minimal pay they get when they cannot work and receive public
assistance. In other words, welfare support is a disincentive to work. Unlike the other
view, individuals are seen as rational man, not inadequate individuals.
What follows is a discussion of institutional factors that make it harder for a
recipient to break free from welfare trap, which are overlooked in the literature. First,

a brief survey of how previous studies looked at participation in welfare programs will
be presented. Then, the relation between welfare participation and household credit
constraints and also the relation between welfare participation and household health
will be investigated.
Who Receives Welfare Support?
The literature on the determinants of welfare participation is not diverse, which allows
some degree of consensus among the researchers. All of these studies develop a model
based on the premise that households try to maximize utility in their decision to
participate in a welfare program (Barr & Hall 1981, Moffitt 1983 and Ashenfelter
1983). In these models, work is naturally irksome and welfare receipt is an alternative to
it. Therefore, all of these studies conclude that the amount of benefits positively and
significantly determine a households participation in a welfare program (Moffitt
1992). Moreover, despite the mixed results on the significance, household wage
appears to share the same negative sign in most of these studies (Table 6 in Moffitt
1992: 20). Of course, there are other non-financial factors that are taken into account
in these studies, such as race, gender, education, health and marital status of the
household heads, but for the sake of the neoclassical argument these studies attempt
to make, these variables are of trivial importance at best.
Although elegant in presentation, there is a practical problem in these models.
Not all eligible households participate in welfare programs. This common aspect of
welfare participation is due to three causes, which are often discussed in the literature.
First, there is the issue of welfare stigma. There is ample evidence to support the
existence of social stigma associated with welfare assistance (Horan and Austin 1974,
Nichols-Casebolt 1986, Jarrett 1996, Ranney & Kushman 1987, Stuber &
Kronebusch 2004 and Stuber & Schlesinger 2006). The fact that some eligible
households chose not to participate in a welfare program, due to perception of welfare
stigma, is partly addressed by Moffitt (1983) in a utility maximizing framework. Yet,
there are other reasons that prevent households from participating in a program for
which they are eligible, such as simple unawareness, the unobserved costs of filing, or
complying with the program. None of these models considered the process of
receiving welfare assistance as a multi-step ordeal, in which awareness, eligibility,
participation, acceptance and enrollment are different stages (Heckman & Smith
2004). Therefore, the result that the amount of benefit in a welfare program is a
positive determinant of participation is a biased one, because the households that
dont participate in a program (dependent variable = 0) are likely to be eligible but
unaware of the program, hence also unaware of the amount of benefits. On the other
hand, participation means awareness of the program, therefore awareness of the
amount of benefits. This is a simple issue of adverse selection, which all of these
models ignore.
Another issue with these models is that they rarely explain how non-financial
factors such as race, gender, health and education, and also some financial factors
other than family income, such as assets and credit constraints, affect the participation

of a household. The main problem, however, is that receiving welfare is seen as an


individual rational choice based on perfect information. In fact, receiving welfare
assistance is predominantly a matter of need rather than a choice.
The table below presents data from 2013 SCF. The year selection is to avoid any
recession-related bias in the empirical findings that might have existed in the 2010
dataset. Also, 2013 is the most recent dataset on US household finances that is
currently available.
Table 1. Characteristics of Welfare Recipients
welfare=0 welfare=1
Total
inonwelfare
95.678
2.260
84.443
asset
995.790
3.716
876.344
age
52.664
45.090
51.753
race
0.770
0.484
0.736
gender
0.799
0.503
0.763
children
0.783
1.260
0.840
married
0.591
0.210
0.545
cohabit
0.068
0.178
0.082
educ
14.228
12.003
13.961
health
3.081
2.530
3.015
college
0.602
0.165
0.549
insurance
0.908
0.882
0.905
emppart
0.661
0.347
0.623
empfull
0.088
0.136
0.093
crationed
0.165
0.450
0.189
creditlimit
4.212
0.155
3.724
Source: 2013 SCF, Federal Reserve
Note: Please see Table 3 at the end for
descriptions of the variables.
On average, welfare recipients (welfare=1) have less before-tax non-welfare income
(incnonwelfare) and less financial assets (asset) for obvious reasons. In order to be
eligible for welfare assistance, a household should have limited resources and income.
To be eligible for TANF cash assistance for instance, household income for a family of
three had to be less than $9,3531 a year on average for 48 contiguous states and the
District of Columbia in 2012. This was roughly 49% of the federal poverty guideline
for a family of three in that year2. Only in Wisconsin, the income requirement, which
was $21,948 per year, was actually above the poverty guideline in 2012 (Falk 2014).
There are also asset limits for households to be eligible for TANF in many states,
except Alabama, Colorado, Louisiana, Maryland, Ohio and Virginia. For the rest of
the states, asset limit ranges from $10,000 in DE to $1,000 in Georgia, Indiana,
Missouri, New Hampshire, Oklahoma, Pennsylvania, Rhode Island, Texas and
1
2

Calculated based on Table A-1 in Falk (2014).


Based on U.S. Department of Health and Human Services (HHS) poverty guidelines in 2012.

Washington. In most states, one vehicle per household is not considered an asset,
whereas in some states, only a portion of the vehicle value is excluded. The house that
the household is residing in is again not considered in the asset limit.
Participants of welfare programs appear to be younger on average (age) and have
more children (children), which are results consistent with the literature and need little
explanation. Similarly, recipients are more likely to be non-white (race=0), and also
female (gender=0). Both of these variables appear to be significant factors increasing the
probability of receiving welfare support in the literature. The reasons for this fact are
often overlooked. This could be an evidence of non-income related obstacles female
and non-white recipients have to face in the welfare participation process (Monnat
2010 & 2010, Keiser et al. 2004 and Gooden 2004).
Another consistent finding is that welfare recipients are less likely to be married
(married=1), but more likely to cohabit (cohabit=1) with a partner. Welfare programs
have neither a clear definition of marital arrangements of households nor
requirements specifically referring to these arrangements. Given the liberty given to
states in managing programs such as TANF and SNAP, the way different marital
arrangements are considered in a program can vary significantly across the states
(Moffitt et al., 2009). Despite this variation, it is still possible to distinguish a
standalone affect of marital status on welfare dependency for two main reasons. First,
TANF has often been argued to create a disincentive for marriage and an incentive to
cohabit with a partner, which is also called Marriage Penalty in the literature
(Moffitt et al. 2009 and Alm et al. 1999). If such an effect exists, married couples will
be less likely to receive benefit or will receive less benefit than the couples that are not
married but living together, or single parent households.
Of course, these descriptive statistics have little explanatory power if any. These
variables can be plugged into a multivariate regression to analyze the isolated effects of
each one of them. However, in order to do so, a concise model for household decision
for welfare participation should be constructed first. As pointed out earlier, this model
should take into account the multi-step process for participation in a welfare program.
Even after doing so, there exists a greater problem that needs to be solved.
Participation in welfare programs has important implications for participants health
and access to credit. And it is the circular causation between welfare participation and
household health, also between welfare participation and household credit constraints
that will be discussed in the following sections.
The Welfare-Credit Nexus
The effects of asset-tested public assistance programs on household saving behavior
have grabbed significant attention in the literature (Ziliak 2003). However, the
repercussions of welfare program participation on households access to credit have
not been a prevalent topic. A few studies mention program participation as a possible
factor that can reduce household access to credit markets (Duca & Rosenthal 1991 &
1993, Duca & Whitesell 1995 and Chakravarty & Scott 1998). These studies use SCF
households as their sample and share similar conclusions on the sign and the

significance of their coefficients. However, a careful look at the questions asked in the
survey would reveal a fundamental mistake in their model selection. The following
question is asked to households in SCF: In the past five years, has a particular lender
or creditor turned down any request you (or your husband/wife/partner) made for
credit, or not given you as much credit as you applied for? If a household answered
yes to this question and either didnt reapply credit or reapplied with no luck, that
household is defined as credit-rationed. A household is defined as a welfare recipient if
they received some sort of public assistance in the survey year. If a household was
credit-rationed in any year preceding the survey year, participating in a welfare
program cant technically be a factor that may have affected their access to credit. But
then, how can we find evidence of such a relation in SCF?
Luckily, SCF conducted a panel survey in 2009 amongst the households that were
surveyed in 2007. The question in 2009 asked: In the past two years, has a particular
lender or creditor turned down any request you (or your husband/wife/partner) made
for credit, or not given you as much credit as you applied for? The way this question
is asked in the panel survey opens the door for some interesting inference. If a
household received welfare assistance in 2007 and was also credit-rationed in 2009,
there might be a statistical relation between these. Looking at the panel survey reveals
that 1,005 of 1,343 welfare recipients in 2007 also received welfare support in 2009. A
household that received welfare support in 2007 is around 30% more likely to be
credit rationed in 2009. This finding suggests an important conclusion. It appears that
welfare recipient households, because of reasons other than their limited income and
assets, have more restricted access to credit than the non-recipient households.
Although the reasons for this are too crucial to overlook, I will leave this discussion for
another study and only ask the following question: Could lending agencies be
discriminating against welfare program participants?
The Welfare-Health Nexus
Studies in the literature show that participation in welfare programs is socially
stigmatized in the society in different ways (Stuber & Kronebusch 2004), at varying
degrees for different programs (Ranney & Kushman 1987) and for different groups of
participants (Jarrett 1996 and Stuber & Schlesinger 2006). Although stigma is
perceived at varying degrees depending on the social situation participant is in, and
participants cultural history, it exists for every welfare recipient (Rogers-Dillon 1995).
The negative effects of the stigmatization on the psychological well being of
participants are documented in various studies (Nichols-Casebolt 1986, OCampo &
Rojas-Smith 1998, Jayakody & Stauffer 2000 and Petterson & Friel 2008). These
negative effects also tend to perpetuate in the post-participation years, further
aggravating recipients socio-economic hardship (Danziger et al. 2008).
The descriptive statistics from the 2013 survey showed that welfare recipients have
lower health status than non-recipients. Even after controlling for other factors in a
multivariate estimation, health status appears to be a significant predictor of program

participation3. However, it is difficult to deduce any causality from this information.


The panel survey, on the other hand, can provide some evidence for causality. 338 of
1,005 households who received welfare support in 2007 left welfare in 2009. These
households have improved health status (2.54) on average compared to those who
stayed on a welfare program (2.49). 610 households that didnt participate in welfare
programs in 2007 but enrolled in 2009 saw a slight worsening of their health status on
average from 2.68 to 2.69. 17,332 households who didnt receive welfare support in
any year had a significantly higher health status than any other group (3.18).
A multivariate regression will provide better support for the isolated impact of
welfare program participation on household health. The results from the estimation of
a simple regression model are presented below. Leaving any concern about the model
selection, fit and reliability aside, the results show one simple thing: participation in
welfare programs decreases the household self-reported health status on average when
other variables are controlled. Another important result is that being credit-rationed
has also a statistically significant negative impact on household health.
Table 2. OLS Estimation Results,
Dependent Variable: Health 09
income09
0.000***
empfull09
0.237***
emppart09
0.180***
race
0.136***
gender
0.089***
children09
-0.001
age09
-0.004***
smoker09
-0.224***
educ09
0.060***
insurance09
-0.104***
welfare07
-0.083
welfare09
-0.149***
crationed09
-0.091***
Constant
2.316***
Observations
7,149
R-squared
0.168
Source: 2007-09 Panel SCF, Federal Reserve
Note: *** p<0.01, ** p<0.05, * p<0.1
Conclusion
Participating in a welfare program and staying in it are often understood as rational
choices households make in order to maximize utility and minimize pain, which is
often associated with work effort. Therefore, it is only second nature to suggest that

The results for this regression are not presented here.

households will choose to stay in a welfare program if they think that it brings them
more utility to receive benefits and not work as much, than not receive benefits
because they worked more and had higher income. According to this line of logic,
welfare trap exists because the welfare system is full of disincentives for households to
increase their work effort in order to keep on receiving benefits.
The findings of this study provide support for a different approach to the issue of
welfare trap. It appears that households access to credit and health, besides many
other important factors, can significantly affect whether or not they will need and
participate in welfare programs. Moreover, households on welfare programs suffer
negative health consequences, even after they stop participating in programs, and also
their access to credit worsens during their welfare spells. Poor health is no doubt a
great obstacle for the poor in their effort to find jobs. In that regard, receiving welfare
support is a self-perpetuating situation for many participants. Likewise, with their
limited access to credit, welfare recipient households will also have restricted access to
crucial resources they can use to improve their health and other socio-economic
conditions such as education.
These findings by no means suggest that the welfare system is what causes the
poor to stay in welfare. The welfare system in the U.S., despite its chronic problems
that need to be addressed, is the only means of survival for many in need. To the
contrary, what causes welfare trap is the way basic necessities are provided in the U.S.
health care, for instance, has been growing to be a predominantly debt-financed
necessity. Although there are programs such as Medicaid to address the health issues
of the poor, the limited coverage it provides and the strict eligibility qualifications are
still leaving a gap to be paid out-of-pocket. This is the point where having a credit card
is the only help many people can get. However, with restricted access to credit, many
welfare recipients cant get that help either.
In order to solve the welfare dependency or welfare trap of the poor, health care
should be made more affordable, especially for the higher risk groups with limited
resources. As long as health care is expensive and requires borrowing to finance, it will
perpetuate the problem. Another necessity is education, which is not discussed in this
study. The findings of this showed that welfare recipients are more than 40% less
likely to have a college degree. Again, there are federal grants for low-income students,
but it is not enough to make higher education a possibility for many participants of
welfare programs. This is also an intergenerational issue. Studies show that children of
welfare recipients receive less education, which is an important factor that creates
intergenerational dependency on welfare (Ku, 2001 and Ku & Plotnick, 2003). Better
health and better education are two fundamental conditions that improve ones
possibility to find a job and keep it. Only one of these conditions is covered in this
study in depth because the topic of education requires a separate study. Without
improving access of the poor to better health and education, any public policy will
remain short of solving the problem of welfare trap.

Table 3. Descriptions of Variables


age
age09

Age of the household head in 2013 (Number of years)


Age of the household head in 2009 (Number of years)

asset

Paper assets of the household in 2013, excluding residential property and vehicles
(10,000$)

children
children09

Number of children in the primary economic unit in 2013


Number of children in the primary economic unit in 2009

cohabit

=1 if household head cohabits with a partner in 2013, =0 otherwise

college

=1 if household head has college degree in 2013, =0 otherwise


=1 if household is turned down for credit or couldnt obtain the desired amount in the
five years prior to 2013, =0 otherwise
=1 if household is turned down for credit or couldnt obtain the desired amount in the
two years prior to 2009, =0 otherwise
The total credit limit on the credit cards the household has in 2013 ($)
Years of education for the household in 2013, averaged for married and cohabiting
couples
Years of education for the household in 2009, averaged for married and cohabiting
couples
=1 if household head is employed full-time in 2013, =0 otherwise
=1 if household head is employed full-time in 2009, =0 otherwise
=1 if household head is employed part-time in 2013, =0 otherwise
=1 if household head is employed part-time in 2009 =0 otherwise
Gender of the household head, =1 if male, =0 otherwise
Health status index (1-4) for the household in 2013, averaged for married and
cohabiting couples
Health status index (1-4) for the household in 2009, averaged for married and
cohabiting couples
Before-tax income of the household from all sources in 2009 (10,000$)
Total income of the household from sources other than welfare assistance in 2013
(10,000$)
=1 if at least one person is covered by some kind of health insurance in the household
in 2013, =0 otherwise
=1 if at least one person is covered by some kind of health insurance in the household
in 2009, =0 otherwise
=1 if household head is living with a spouse in 2013, =0 otherwise
Race of the household head, =1 if white, =0 otherwise
=1 if either household head or the spouse/partner was a smoker in 2009, =0 otherwise
=1 if household received welfare assistance such as TANF, SSI or SNAP in 2013, =0
otherwise
=1 if household received welfare assistance such as TANF, SSI or SNAP in 2007, =0
otherwise

crationed
crationed09
creditlimit
educ
educ09
empfull
empfull09
emppart
emppart09
gender
health
health09
income09
inonwelfare
insurance
insurance09
married
race
smoker09
welfare
welfare07
welfare09

=1 if household received welfare assistance such as TANF, SSI or SNAP in 2009, =0


otherwise

References
Alm, James, Stacy Dickert-Conlin and Leslie A. Whittington. Policy Watch: The Marriage Penalty. The
Journal of Economic Perspectives 13, 3 (1999): 193-204.
Ashenfelter, Orley. Determining Participation in Income-Tested Social Programs. Journal of the American
Statistical Association 78, 383 (1983): 517-525.
Barr, N. A. and R. E. Hall. The Probability of Dependence on Public Assistance. Economica, New Series,
48, 190 (1981): 109-123.
Chakravarty, Sugato and James S. Scot. Relationships and Rationing in Consumer Loans. The Journal of
Business 72, 4 (1999): 523-544.
Danziger, Sandra K., Maroia J. Carlson and Julia R. Henly. Post-Welfare Employment and Psychological
Well-Being Women & Health 32, 1-2 (2001): 47-78.
Duca, John V. and Stuart S. Rosenthal. An Econometric Analysis of Borrowin and Household Debt.
Research Paper 9111, Federal Reserve Bank of Dallas, 1991.
Duca, John V. and Stuart S. Rosenthal. Borrowing Constraints, Household Debt, and Racial
Discrimination in Loan Markets. Journal of Financial Intermediation 3, 1 (1993): 77-103.
Duca, John V. and William C. Whitesell. Credit Cards and Money Demand: A Cross-sectional Study.
Journal of Money, Credit and Banking 27, 2 (1995): 604-623.
Falk, Gene Temporary Assistance for Needy Families (TANF): Eligibility and Benefit Amounts in State
TANF Cash Assistance Programs., Research Paper, Congressional Research Service, 2014.
Fraser, Nancy and Linda Gordon A Genealogy of Dependency: Tracing A Keyword of the U.S. Welfare
State. Signs 19, 2 (1994): 309-336.
Gooden, Susan Tinsley. Examining the implementation of welfare reform by race: Do blacks, hispanics and
whites report similar experiences with welfare agencies? The Review of Black Political Economy 32, 2 (2004):
27:53.
Heckman, James J. and Smith, Jeffrey, A. (2004) The Determinants of Participation in a Social Program:
Evidence from a Prototypical Job Training Program, Journal of Labor Economics, Vol. 22, No. 2, pp. 243-298.
Horan, Patrick M. and Austin, Patricia Lee (1974) The Social Bases of Welfare Stigma, Social Problems,
Vol. 21, No. 5, pp. 648-657.
Jarrett, Robin L. (1996) Welfare Stigma among Low-Income, African American Single Mothers, Family
Relations, Vol. 45, No. 4, pp. 368-374.
Jayakody, Rukmalie and Dawn Stauffer. Mental Health Problems Among Single Mothers: Implications for
Work and Welfare Reform. Journal of Social Issues 56, 4 (2000): 617-634.
Keiser, Lael R., Peter R. Mueser and Seung-Whan Choi. Race, Bureucratic Discretion, and the
Implementation of Welfare Reform. American Journal of Political Science 48, 2 (2004): 314-327.
Ku, Inhoe. The Effect of Welfare on Childrens Education. Social Service Review 75, 2 (2001): 245-270.
Ku, Inhoe and Robert Plotnick. Do children from welfare families obtain less education? Demography 40, 1
(2003): 151-170.
Moffitt, Robert. An Economic Model of Welfare Stigma. The American Economic Review 73, 5 (1983):
1023-1035.
Moffitt, Robert. Incentive Effects of the U.S. Welfare System: A Review. Journal of Economic Literature30, 1
(1992): 1-61.
Moffitt, Robert. Beyond Single Mothers: Cohabitation and Marriage in the AFDC Program. Demography
35, 3 (1998): 259-278.
Moffitt, Robert. The Role of Nonfinancial Factors in Exit and Entry in the TANF Program. The Journal of
Human Resources 38, Special Issue on Income Volatility and Implications for Food Assistance Programs
(2003): 1221-1254.
Moffitt, Robert A., Robert T. Reville, Anne E. Winkler and Jane McClure Burstain. Cohabitation and
Marriage Rules in State TANF Programs. HHS Research Paper, 2009.
Monnat, Shannon M. Toward a Critical Understanding of Gendered color-Blind Racism Within the U.S.
Welfare Institutions. Journal of Black Studies 40, 4 (2010): 637-652.
Monnat, Shannon M. The Color of Welfare Sanctioning: Exploring the Individual and Contextual Roles
of Race on TANF Case Closures and Benefit Reductions. The Sociological Quarterly 51, 4 (2010): 678-707.

Nichols-Casebolt, Ann. The Psychological Effects of Income Testing Income-Support Benefits. Social
Service Review 60, 2 (1986): 287-302.
Patricia, OCampo and Lucia Rojas-Smith. Welfare Reform and Womens Health Review of the Literature
and Implications for State Policy. Journal of Public Health Policy 19, 4 (1998): 420-446.
Petterson, Stephen M. and Lisa V. Friel. Psychological Distress, Hopelessness and Welfare. Women &
Health 32, 1-2 (2001): 79-99.
Ranney, Christine K. and John E. Kushman. Cash Equivalence, Welfare Stigma and Food Stamps.
Southern Economic Journal 53, 4 (1987): 1011-1027.
Rogers-Dillon, Robin. The Dynamics of Welfare Stigma. Qualitative Sociology 18, 4 (1995): 439-456.
Stuber, Jennifer and Karl Kronebusch. Determinants of Participation in TANF and Medicaid. Journal of
Policy Analysis and Management 23, 3 (2004): 509-530.
Stuber, Jennifer and Mark Schlesinger. Source of stigma for means-tested government programs. Social
Science and Medicine 63, 4 (2006): 933945.
Ziliak, James P. Income Transfers and Assets of the Poor. The Review of Economics and Statistics 85, 1
(2003): 63-76.